with-agents

Keynote: Linus Torvalds, Creator of Linux & Git, in Conversation with Dirk Hohndel

Open Source Summit

Linus Torvalds speaks with Dirk Hohndel about Linux development and his view of Artificial Intelligence.

AI's Impact on Open Source Infrastructure

  • Infrastructure Disruption: Linus says AI has been "very disruptive" to kernel infrastructure. AI crawlers scanning the kernel.org source code have strained their systems and caused problems.
  • Spam & Low-Quality Reports: There is a rising issue with "made up" bug and security reports generated by people misusing AI tools. These false positives waste maintainer resources, though Linus notes it hasn't been as severe for the Linux kernel as it has been for other projects like curl.

"Vibe Coding" & AI as a Tool

  • Skepticism for Real Products: Linus is skeptical of "vibe coding" (using AI to write code based on "vibes" or natural language prompts) for serious product maintenance. He believes that while it might get you 90% of the way there, the last 10%—the hard part of maintenance—will likely prove that this approach isn't viable for long-term projects.
  • Useful for Learning: Conversely, he views AI coding tools as a positive gateway for beginners. He says they lower the barrier to entry, allowing people to create working programs and become interested in computers more easily than before.
  • Current Utility: For the kernel specifically, AI isn't quite "there yet" for complex code generation. It is currently being used experimentally to help maintainers with tasks like backporting patches, but it remains a tool rather than a replacement.

The Future of Software Jobs

  • Evolution, Not Extinction: When asked about layoffs and the fear of AI replacing developers, Linus compares AI to previous innovations like compilers. Just as compilers freed programmers from writing assembly by hand but didn't eliminate the need for programmers, he believes AI will simply be another tool that removes minutia.
  • Productivity & New Opportunities: His "gut feel" is that AI will make programmers more productive and open new areas of development. He says this could lead to more software jobs rather than fewer.

Hardware & Industry Shifts

  • Nvidia & Linux: One positive side effect of the AI boom is that it forced Nvidia to become a "good player" in the Linux space. Since Linux is critical for AI cloud infrastructure, Nvidia now engages more deeply and constructively with the kernel community than they did 20 years ago.
  • General Purpose vs. AI Hardware: Despite the hype around AI hardware (GPUs/NPUs), Linus maintains that the general-purpose CPU is still the most interesting part to him. He views the AI workload as just another application running on top of the kernel, similar to how big databases or commercial apps have always run on Linux.

Full video: https://www.youtube.com/watch?v=tWx769t1JKg(opens in a new tab)